Snowbird writes:

What a brilliant deduction.

Thank you.

I suppose next you will then postulate that the
simulator has a more accurate flight dynamics model than the real airplane.

The real airplane is not a model, so this statement has no meaning.

Then tell me why it's self-evident that a table-driven flight dynamics model
would always be better than a real-time differential equation-driven.

Not always, but usually, especially cost-wise. The idea of a simulator is to
simulate a real aircraft, not real flight. It's more important that the
behavior of the sim match the real aircraft being simulated than it is for the
sim to approximate real flight in all regimes.

Full-motion ATP sims handle spins and other unusual attitudes poorly, but
since they are not used to simulate those unusual attitudes, it doesn't
matter, especially since they simulate normal flight with extraordinary
accuracy.